Evaluating the evidence for treatments is a vital pursuit. Research is occurring all over the world all the time to evaluate treatments for chronic pain. Some of the concerns of consumers of products include how to judge the quality of the research. To answer this concern, researchers have compiled a comprehensive library of systematic reviews and meta-analyses which summarise and report on the findings of high quality medical and other healthcare research. This is called the Cochrane Reviews. Where possible, we will refer to these refers as we journey through this Chapter looking at medicines and other treatments for pain.

A simple way to communicate the risks and benefits of a treatment is to use the Numbers Needed to Treat (NNT) vs Numbers Needed to Harm (NNH) concept.

The Number of people Needed to Treat is the number of people receiving a treatment who get a reduction of their pain by half (50%) compared to a placebo or no treatment. For example if every person that walks into the doctors surgery is provided treatment ‘X’ and they all get the 50% reduction in their pain, the NNT=1. This is the ideal NNT. It assumes that it works for every person taking the treatment and that the control does not.

A NNT of 2 is saying that 1 person in 2 gets a 50% relief of their pain. NNT of 4 means that 1 in 4 people will get that 50% benefit, 3 will not. Clearly, the lower the NNT, the better the treatment.
The Number Needed to Harm (NNH) is another good way of evaluating a treatment or medicine. It is the number of people who get harmed by a treatment. So a NNH of 1 means that everyone gets harmed. A substance like rat poison might be a good example of a NNH=1! So NNH=2 is where 1 in 2 is harmed. Obviously the higher the NNH the better, fewer people get harmed by the treatment.
The NNT vs NNH concept is used extensively by pharmaceutical benefits agencies, where if the NNH is very low and the NNT is very high, the benefit might not be seen to be enough to warrant the support of the medicine.

The following table is sourced from the Fremantle Hospital Pain Medicine Unit information for patients. More examples can be found at http://www.medicine.ox.ac.uk/bandolier/band50/b50-8.html



(morphine, oxycontin, buprenorphine, fentanyl)

Neuropathic pain

(no studies longer than 3 months)


4.2 (nausea, constipation)

7.1 (dizziness, vomiting)

Tramadol (Tramal, Zydol, Durotram) Neuropathic pain 3.8 8.3
Amitryptyline (Tryptanol, Endep) Neuropathic pain 3.6 28 (major) 6 (minor)

Gabapentin (Neurontin)

Pregabalin (Lyrica)

Chronic pain

Diabetic neuropathy

Post-herpetic neuralgia




3.7 (minor)

Venlafaxine (Efexor)

Duloxetine (Cymbalta)

Neuropathic pain 3.1 16.2 (major) 9.6 (minor)

Panadol (4 g/day)

Panadol Slow Release

Arthritis pain 4-5 12